learning algorithm 0.002171319
new algorithm 0.0016950959999999999
ssl algorithm 0.0016546339999999999
data set 0.001428115
first set 0.00137322
objective function 0.001294595
algorithm 0.0012943
learning problem 0.0012850510000000002
training data 0.0012837880000000001
weighted graph 0.0012610759999999999
ssl graph 0.001244188
large data 0.0012335089999999998
loss function 0.001217517
machine learning 0.001213422
ing set 0.001204715
graph regularizer 0.001197544
supervised learning 0.001185336
transductive learning 0.001182514
feature vectors 0.001174748
learning framework 0.001160424
ing data 0.00115631
unsupervised learning 0.001155788
graph regions 0.001155595
undirected graph 0.001153166
graph structure 0.0011435899999999999
graph transduction 0.001135039
graph trans 0.001111646
data point 0.001110502
graph vertex 0.001109799
graph construction 0.001108751
spectral graph 0.00110713
test set 0.001096734
tropy function 0.001094039
learning proce 0.001093667
probability distribution 0.001090248
unlabeled data 0.0010791239999999999
other words 0.001065459
loss model 0.00106489
data sets 0.001048795
first similarity 0.001041517
labeled data 0.001040346
convex set 0.001036475
input data 0.001028949
data points 0.001020219
possible method 0.001015615
training point 0.0010145800000000002
labeled distribution 0.0010035719999999999
first term 0.001002093
webkb set 9.99053E-4
label training 9.97211E-4
other information 9.95055E-4
state convergence 9.9496E-4
transduction set 9.89445E-4
other approaches 9.681049999999999E-4
reuters data 9.580159999999999E-4
topic categories 9.55976E-4
tire set 9.54135E-4
optimization parameter 9.52965E-4
large number 9.52745E-4
webkb data 9.506479999999999E-4
labeled training 9.44424E-4
form solution 9.38291E-4
uniform distribution 9.31525E-4
closed data 9.28294E-4
unseen data 9.269069999999999E-4
valid distribution 9.263170000000001E-4
similar updates 9.20297E-4
data sample 9.159799999999999E-4
same solution 9.1384E-4
input features 9.104639999999999E-4
data uncertainty 9.070879999999999E-4
simple features 9.034659999999999E-4
supervised training 9.022500000000001E-4
feature 9.01796E-4
many problems 9.015259999999999E-4
document classification 9.00734E-4
classification results 8.90638E-4
first transduction 8.86145E-4
graph 8.83854E-4
other hand 8.77923E-4
function 8.77349E-4
learning 8.77019E-4
such category 8.76478E-4
other universities 8.74635E-4
ssl algorithms 8.718350000000001E-4
agglomerative distribution 8.69953E-4
point methods 8.68125E-4
tial topic 8.595829999999999E-4
first polychotomy 8.52607E-4
such modification 8.376540000000001E-4
such situations 8.35328E-4
optimization problem 8.30028E-4
proposed approach 8.2719E-4
ing ssl 8.267890000000001E-4
final labels 8.261049999999999E-4
current state 8.20566E-4
form solutions 8.18666E-4
new objective 8.180419999999999E-4
important class 8.166919999999999E-4
classification problems 8.166499999999999E-4
